This project seeks to create, implement, and evaluate the efficacy of a 6- month therapeutic community (TC) treatment program for inmates who have a history of substance abuse and who are going from prison to a work release center prior to release from custody. All subjects (N=620) will have a history of serious drug abuse (IV drugs, cocaine/crack, PCP, methamphetamine) in the year prior to incarceration but no history of serious violent crimes. The primary evaluation compares: 1) the intensive TC work release condition (n=360), to 2) the conventional work release approach (n=260). A secondary evaluation compares two groups of subjects in the TC work release condition: 1) inmates who have been in an in-prison TC prior to work release (n=100), and 2) inmates who have no in-prison TC experience (n=260). Subjects will be randomly recruited into the study from among inmates in the appropriate pools who are classified as eligible for work release. Subjects in all 3 conditions will receive urine and HIV testing as well as an extensive questionnaire (incorporating key elements of the Addiction Severity Index and the AIDS Initial Assessment, as well as background and psychosocial variables) immediately prior to their classification to work release. They will be retested and reinterviewed at completion of work release and subsequently at 6 months and 12 months, resulting in four waves of data. The target sample will be 50% white, 50% black, 80% male, and 20% female. Data will be used to empirically test the relative effectiveness of the TC work release center as compared to conventional work release, and to test the relative effectiveness of TC work release for those with prior TC experience and those with no prior experience. Criteria for the comparisons will make use of both self- report and physical test data. These criteria include both absolute (yes/no) and time relative (how long before) measures of drug use, other criminal activity, and recidivism. In addition, analyses will seek to identify etiological models for these criteria that consider additional variables beyond the treatment condition, including background and psychosocial indicators. Analysis techniques to be employed include cross-tabulations with adjusted repeated tests of significance, discriminant function analysis, structural equation models for longitudinal data, and modified event-history analysis. Results of his research will provide important information on the use of TC treatment in a setting where it can have a maximum impact on the drug-using behavior of a group at particularly high risk for drug relapse, HIV infection, criminal activity, and recidivism. The success of the TC work release program would provide a cost-effective benefit to public health, corrections, and the general social welfare. This study fully complies with federal requirements for research involving prisoners as human subjects.